IBM Israel
Skip to main content
 
Search IBM Research
   Home  |  Products & services  |  Support & downloads  |  My account
Select a Country Select a country
IBM Research Home IBM Research Home
IBM Haifa Labs Homepage IBM Haifa Labs Home
IBM Haifa Labs Homepage IBM Haifa Labs Leadership Seminars

Technologies in an On-Demand World 2003

Invitation
The Program
Abstracts
Visitors information
Confirmed participants
Feedback


Abstracts
The Application Perspective in the On-Demand World (Abstract)
Shay Naeh, Director APM Product Management, Veritas

Applications become more and more complex. Today's applications are mission-critical multi-tier applications, supporting hundreds to thousands of transactions and users, which are located at different remote locations.

Understanding the application behavior and performance patterns from a business perspective as well as from the underlying IT resource usage perspective is challenging.

Without "application knowledge" we cannot understand on-demand infrastructure requirements.

Veritas, through the acquisition of Precise Software, has developed the Veritas Ił suite, which constantly, 7x24, monitors business applications to detect service level (SLA) breaches as well as application behavior deviations, find their root cause across the application infrastructure and suggests remedies. Application collected metrics are kept in a Performance Warehouse, which is a long-term repository used for trend and exception analysis.

This session discusses different application and infrastructure related aspects in the new on-demand paradigm.


Active Monitoring for Straight-Through-Processing in Financial Markets (Abstract)
Tali Yatzkar-Haham, IBM Haifa Research Labs

Periodically, every industry undergoes a major change in the methods it employs to satisfy customer and market demands. The financial services industry is presently undergoing such a change, known as straight through processing (STP). STP now encompasses a revised organizational, process, and technology model for financial industry organizations in all segments and all geographies. Straight through processing can be viewed as a series of uninterrupted processes across and throughout an enterprise.

Clients moving into STP will be replacing fragmented batch based processes with a continuous process operating in near real-time and spanning multiple organizations. To adequately manage in this new environment, securities operations will need to be able to monitor the business processes, the potential problems arising from unexpected buildup of work requiring the involvement of clerks, and monitor delays in delivery of information from external sources. STP needs middleware able to support the complexity of continuous processing of external event triggers, near as possible real-time performance.

Our Transaction Monitoring application comes to serve as a monitoring solution to STP transaction processing. We use the active technology tools, Amit and Adi to monitor transactions effectively, through their process flow, from inception to settlement. Our application (a) provides overall status, statistics and deviation from key business indicators, (b) enables queries on on-going and completed transactions, and (c) alerts on high level business exceptions and anticipated problems.


Telephone Services: Science, Engineering and Management (Abstract)
Prof. Avishai Mandelbaum, Technion

The world of call centers is vast, and growing quickly. It is estimated that 3% of the USA workforce is involved with call centers and that the industry is expanding at an annual rate of 20%. Telephone services are unparalleled in scope, service quality and operational efficiency. For example, in a large best-practice call center, hundreds of agents attend to thousands of callers per hour, with an average occupancy of 95%, about half of the callers being answered immediately, and the rest delayed for scarcely few seconds.

The design of call center operations, and the management of their performance, has traditionally relied on classical queuing theory. However, the modern call center and its emerging successor the contact-center (= telephone + IVR + internet + email + chat + ...), are challenging the relevance of this conservative approach. My goal is to survey ongoing research that addresses some of these challenges, including skills-based-routing and square-root safety-staffing.

Some of the research has been taught at the Technion in a course on "Service Engineering". Course material is downloadable from: http://ie.technion.ac.il/serveng2003.
The "References" menu on this website has copies of most of the papers that support my lecture.


The Value of Information in Shared Contexts (Abstract)
Prof. Sheizaf Rafaeli, BGS, Director of the Center for the Study of the Information Society, HU

In this talk I will discuss the subjective value of information. How much is information worth to you? What would predict the propensity or agreement to share information over networks?

In a networked world, much of the competitive advantage of groups and teams depends on having an information edge (knowing more), and on having sharing practices and processes. Sharing processes are made possible and prevalent increasingly through networks, forums, email, distributions, peer-to-peer and the like. The practice and preference to share information remains enigmatic and problematic. What do we know about the individual, group and network level parameters of sharing information?


Active Integration and Self-Stabilization, Case Study (Abstract)
Gil Nechushtai, IBM Haifa Research Lab

Active integration (sometimes called real-time enterprise or zero latency enterprise) deals with situations where some event occurring in the operational system triggers a time constrained decision process, which in turn affects the operational system. Such applications, involving direct customer contact, different business processes, several degrees of automation, and multiple integration problems.

The Active Integration initiative is motivated by the need to provide the appropriate integration from OLTP through to data warehouse analytics in a way that exhibits much less latency than is currently available. This vision is now being realized by technology that is under development in IBM's product and research labs.

The idea of stabilization has been raised in the context of fault tolerance, defining a system as self-stabilizing if regardless of its initial state, it is guaranteed to arrive at a legitimate state in a finite number of steps. The implementation of this idea employs a reflective approach, according to which the system examines the information available about the type of problem, the reason for inability to achieve a consistent state and a solution strategy defined for this case. Using this approach, the system designer does not need to define an exception handler. Definitions of general policies and (if required) a method for collaboration with other systems and users are sufficient for the system to complete the stabilization task.

This talk will cover the Active integration and self stabilization concept through a case study of handling changes in real time for a "Traffic Light" business.


e-business on demand: A Technology Perspective (Abstract)
Fulvio Capogrosso, Distinguished Engineer, VP IBM Academy of Technology

'e-business on demand' is the name IBM has chosen to describe a new movement in business and computing, one that is changing the way technology is deployed and used in business. The dynamics of the market are changing, the next phase of e-business is coming, where companies move beyond simply integrating their various processes-to a world in which they need to be able to sense and respond to fluctuating market conditions in real time. This requires a new way of thinking about business processes and the information technology infrastructure that supports them.

This lecture will address the full content of the IBM 'e-business on demand' proposition, indicate the steps required to become an 'on demand' business, and provide details on some key enabling technologies, such as autonomic computing and grid computing.


Protection on-demand: Ensuring Resource Availability (Abstract)
Prof. Yehuda Afek, TAU, CTO, Riverhead Networks

Current solutions to flooding attacks such as DDoS and Worm attacks either focus on detection (such as intrusion detection systems), or they employ primitive, indiscriminate filtering techniques that block good traffic as well as bad, effectively completing the hacker's job.

A true cyber attack protection must provide both detection and mitigation services that preserve business integrity. This talk introduces a new attack mitigation technology that ensures continued service availability even while under DDoS attacks by combining unique on-demand attack traffic "diversion" techniques with a mechanism that filters "bad" (illegitimate) packets while forwarding "good" (legitimate) packets to the intended server. This method minimizes demand on router resources and does not introduce additional elements onto the normal data path, minimizing the impact on normal network operation or performance.

This unique method cleans malicious traffic at both the ISP and data center levels as well as at the entrance to large enterprises. Various configurations and scenarios for DDoS protection will be presented and discussed.


On-Demand Optimization According to Business Objectives (Abstract)
Dr. Natalia Razinkov, IBM Haifa Research Lab

A central challenge in the runtime management of computing environments in any enterprise is both the initial optimization of these environments, and keeping these environments optimized when changes occur. Current goals for such optimizations typically focus on IT measures (e.g., increasing the site availability by 1%, reducing overall response times to an average of three seconds). However, nowadays, many enterprises are keen on satisfying business objectives, such as maximizing the total income generated by the infrastructure.

Optimizing the IT infrastructure according to such business objectives is not a trivial task, as it is unclear how the setting of parameters at the IT level will affect the business objectives. Moreover, optimizing the IT infrastructure is not a one time effort, as there may be changes that occur in the environment in which the infrastructure operates, which may render any pre-defined setting suboptimal. Two examples of such changes might be failures of hardware and software components, and significant changes in the mix or load of users of this infrastructure. Thus, the solution requires both an automatic mechanism for carrying out IT optimization according to business metrics and an automatic mechanism for recognizing significant changes and re-optimizing the system as a result of such changes, enabling what we term on demand optimization.


eBusiness Survival Kit (Abstract)
Rony Zarom, Founder and CEO, Etagon

The struggle for survival has never been tougher in the enterprise marketplace. Competition and harsh economic conditions have forced companies to constantly evolve and adapt to changing market conditions. Survival of the fittest depends on achieving the right balance between aggressive and defensive approaches. On one hand, enterprises must react aggressively to changing market needs; while on the other hand, they are bound by the limitations of their IT infrastructure that draw them into a more defensive approach.

Etagon, an Israeli company, has developed software to easily cope with the dynamics of deploying and managing eBusiness data-centric application clusters on blade server solutions. Targeting the enterprise, the company's Centergon application management system (AMS) provides a simplified way to integrate, configure and manage IT services and applications in a data center. The first Centergon application module to be released, ManageOn, is specifically designed to simplify the deployment, configuration and management of the Oracle9i Real Application Clusters database. ManageOn enables any database administrator to deploy a multi-node Oracle9i RAC database on an IBM BladeCenter in less than one hour, despite the complexities usually associated with such a deployment. In addition, ManageOn provides greater availability and scalability for Oracle 9i RAC by utlizing Centergon's networking and hardware management capabilities. Combined with the latest IBM BladeCenter technology, Etagon software provides a state-of-the-art deployment and management solution for cluster applications in the eBusiness data center.




































  About IBM  |  Privacy  |  Terms of use  |  Contact